Touch pads have been in use for a number of years. Most of them use capacitive sensing to detect when and where the surface of the touch pad has been touched by the user. Flat touch pads typically have an array of many such capacitive sensors, and can infer motion gestures as the user's fingertip drags across the surface and activates more sensors in the array. But if the surface has a complex curvature, it may be more cost effective to place a small number of comparatively large sensors at or under the surface. However, a smaller number of sensors makes the device relatively insensitive to small motions, since the resolution is inherently poor. Further, as devices become smaller, the user is more likely to use his thumb to activate the touch surface. The mechanics of the human thumb are such that extension of the thumb may cause the tip of the thumb to lift off the touch surface as the more proximal portion of the thumb makes contact, leaving the contact area on the touch surface relatively unchanged when the thumb is extended. This could be incorrectly sensed by the device as no motion by the thumb.